Search results

Issue created for Rollback of PR #41399: [Grappler] Fix bug in UpdateSqueezeDims

Status: Open.

#117337 In tensorflow/tensorflow;

· github-actions opened on Apr 30, 2026

tf.debugging.check_numerics silently passes NaN/Inf under jit_compile=True

type:bugBugBug

Status: Open.

#117319 In tensorflow/tensorflow;

· wuyii8941 opened on Apr 30, 2026

tf.nn.batch_normalization returns wrong result under jit_compile=True for large-magnitude inputs

type:bugBugBug

Status: Open.

#117318 In tensorflow/tensorflow;

· wuyii8941 opened on Apr 30, 2026

tf.math.top_k NaN handling is position-dependent in eager, differs from XLA

2.20.0tensorflow 2.20.0tensorflow 2.20.0

comp:xlaXLAXLA

type:bugBugBug

Status: Open.

#117287 In tensorflow/tensorflow;

· wuyii8941 opened on Apr 30, 2026

tf.nn.moments returns wrong variance under jit_compile=True for large-magnitude inputs

2.20.0tensorflow 2.20.0tensorflow 2.20.0

comp:xlaXLAXLA

type:bugBugBug

Status: Open.

#117286 In tensorflow/tensorflow;

· wuyii8941 opened on Apr 30, 2026

tf.image.non_max_suppression with empty inputs crashes the process under XLA jit_compile=True

2.21.0tensorflow 2.21.0tensorflow 2.21.0

comp:xlaXLAXLA

type:bugBugBug

Status: Open.

#117245 In tensorflow/tensorflow;

· rookieLiu2018 opened on Apr 29, 2026

tf.image.adjust_hue suppresses NaNs on GPU but propagates them on CPU

2.21.0tensorflow 2.21.0tensorflow 2.21.0

comp:opsOPs related issuesOPs related issues

type:bugBugBug

Status: Open.

#117189 In tensorflow/tensorflow;

· beanduan22 opened on Apr 28, 2026

python 3,14 support

type:featureFeature requestsFeature requests

Status: Open.

#117172 In tensorflow/tensorflow;

· jdfoxito opened on Apr 28, 2026

Address TODO: respect compiled trainable state in TensorFlowTrainer

awaiting PR mergeawaiting PR mergeawaiting PR merge

comp:kerasKeras related issuesKeras related issues

type:bugBugBug

Status: Open.

#117169 In tensorflow/tensorflow;

· andersendsa opened on Apr 28, 2026

XLA-compiled tf.function returns different result from eager for int8 multiply followed by conditional abs

2.21.0tensorflow 2.21.0tensorflow 2.21.0

comp:xlaXLAXLA

type:bugBugBug

Status: Open.

#117142 In tensorflow/tensorflow;

· rookieLiu2018 opened on Apr 28, 2026

PTQ INT8 via TFLiteConverter — encoder-decoder seq2seq model loses encoder context entirely after conversion

comp:liteTF Lite related issuesTF Lite related issues

TF 2.13For issues related to Tensorflow 2.13For issues related to Tensorflow 2.13

TFLiteConverterFor issues related to TFLite converterFor issues related to TFLite converter

Status: Open.

#117052 In tensorflow/tensorflow;

· AminLO-ma opened on Apr 27, 2026

CPU XLA silently drops NaN in reduce_max / reduce_min / maximum / minimum

2.21.0tensorflow 2.21.0tensorflow 2.21.0

awaiting PR mergeawaiting PR mergeawaiting PR merge

comp:xlaXLAXLA

type:bugBugBug

Status: Open.

#116992 In tensorflow/tensorflow;

· wuyii8941 opened on Apr 27, 2026